Generative AI landed in higher education before anyone had time to prepare. No matter how you feel about these tools—skeptical, overwhelmed, curious, or excited—they’ve created real challenges for faculty. This page brings together what I’ve learned about GenAI, its possibilities, its limits, and its ethics, with the hope that it helps you in ways that make sense for your courses and your comfort level.
Our library has put together an AI toolkit, which includes how AI works and how to assess its content.
“I want to minimize or avoid AI in my courses.”
AI-Cautious Resources
Generative AI has arrived at a moment when faculty are already balancing heavy demands and it’s understandable to feel wary; this section focuses on practical ways to preserve your teaching values and the integrity of student work without needing to overhaul your courses.
Our library has put together a AI resource toolkit, including what AI is and how to assess it.
Deciding on your AI policy
Barnard has some useful infographics on how to start articulating your stance on AI use in your courses. Below the policy infographic is an infographic on how to start navigating concerns on AI use in your assignments.
Designing “AI-resistant” assignments
Start by considering your course learning outcomes (what do you want your students to be able to do by the end of the course?). How might a student demonstrate to you that they’ve achieved that outcome or skill?
As you brainstorm, you can look at the resources below for the kinds of assessments and assignments that are more AI-resistant.
Barnard also has a useful section on their website under the heading “Consider assignment design in the context of ChatGPT“. Some points they make:
- Avoid redesigning assignments solely to prevent AI use; instead consider adapting assignments to emphasize higher-order thinking, ‘process’ work, and verifiable sources (areas where AI tools struggle due to inaccuracies and limited analytical capabilities). Includes an article on adapting college writing assignments.
- While in-class, hand-written assignments may seem tempting, keep accessibility in mind. How do timed, hand-written exams potentially disadvantage students with learning disabilities or whose first language is not English? Further rumination on timed exams in the AI-age: Should we get students back in exam halls?
- Effective strategies include scaffolding assignments with low-stakes practice, incorporating recent texts (which older AI models can’t access), building in peer-review feedback & revision opportunities
From the University of Chicago, their teaching center acknowledges that no assignment can be completely “AI-proof”. Instead they advocate focusing on the learning process and offer practical ways to design coursework that discourages students from relying on GenAI. These approaches shift focus from polished final text to meaningful learning and individual intellectual work, strategies they mention include:
- Use thoughtful retrieval practice: ask students to recall, explain, or apply concepts in their own words rather than rely on summary.
- Break large tasks into stages: require planning, drafts, reflection, and revision so you can assess student development and choices.
- Include reflection on thinking: have students explain how they approached a problem and why they made decisions (process over product).
- Design prompts around personal or course-specific contexts: ask for connections to class discussions, texts, or experiences that AI won’t know.
- Incorporate multimodal work: such as concept maps, visual representations, or collaborative annotation that emphasize active engagement.
Oral exams
Some of our colleagues have also been experimenting with incorporating oral exams. Hamilton College’s teaching center offers faculty resources aimed at designing, supporting, and assessing oral communication tasks. Here are some of their examples of strategies that are meant to make oral communication assignments more transparent, equitable, and effective as demonstrations of student learning:
- Use oral exams purposefully: Clarify whether the goal is to assess communication fluency (e.g., in language courses) or students’ understanding of course content through spoken responses.
- Structure oral exams clearly: Provide a set of potential questions students can prepare for, and use standardized protocols (e.g., random selection from prepared questions) to promote equity and reduce anxiety.
- (from me): If instructors are concerned about students relying on AI-generated drafts, they can explain that exam questions will be closely related but not identical to the shared prompts, reinforcing that strong preparation means grasping the underlying ideas rather than rehearsing fixed answers.
- Gradually deepen complexity: Start with simpler questions and move to more complex follow-ups to scaffold students’ thinking during oral assessment.
- Help students prepare: Give clear expectations, encourage thinking through responses rather than memorizing, and consider tools such as limited reference cards or practice sessions.
- Encourage use of resources: Students can work with peers or TAs to practice responses and build confidence before oral assessments.
Self-assessments
Depending on your learning goals, might a self-assessment work? While there are not “AI-proof” assignments, some prompts may make AI use seem irrelevant or pointless. Some ideas from CSU teaching center:
- Assignment wrappers– Write an analysis of your performance on the last exam/ assignment. What did you do well? What didn’t you do well? How will you improve?
- Learning Journeys (longer assessment): Use five of your low-stakes assessments this semester (quizzes, discussion board posts, chat comments, group work assignments, etc) to describe your learning journey this semester.
- Questions for Discussion Boards:
- Which discussion board post or in-class comment by a peer did you find most thought-provoking? What did you learn from it? Review the syllabus and your course notes. Which discussion or class period was most helpful to you this semester? How did it change your thinking?
- Review the course lectures or readings. Which one changed your understanding of the course material the most?
- Alternative Exam Questions (from a mathematician):
- What mathematical ideas are you curious to know more about as a result of taking this class?Take one homework problem you have worked on this semester that you struggled to understand and solve, and explain how the struggle itself was valuable.
Sample syllabi language
This document houses a wide range of syllabi language, including language prohibiting GenAI use and the instructor’s reasoning (e.g., AI is discriminatory, AI’s material is plagiarism). Search in the document for word stems like “prohibit”, “bias”, “discrim”, “plagiarism” to find specific examples.
AI-Curious Resources
GenAI is often labeled as a world-upending tool; my advice is to start small and ground yourself in the basics. Think about your usual frameworks, such as Backward Design, Bloom’s Taxonomy, or whatever guides your teaching. Instead of asking, “How can I add AI to my course?” begin with your existing learning goals and ask “how might AI support or strengthen this?”
Getting Started: Learning about GenAI, crafting prompts, and more
Begin trying GenAI out for yourself. AI efficiency gains come when you use it daily for multiple weeks in a row.
Try this quick reflection.
- Think of the different parts of your job and how much time they take.
- Identify parts of your work that truly require you — your expertise, judgment, warmth, creativity, or presence.
- Note the things that routinely pull you away from those high-value areas.
With that in mind, ask yourself: Are there tasks on that that third “gets in my way” list where AI could save you time or lighten the load? So you can spend more of your energy on the work only you can do? Examples:
- Softening the tone of my feedback to students or colleagues. When my mind is too tired to be diplomatic, how I craft my prompt to ChatGPT: 1) I describe what I want to communicate to a student/colleague in my unorganized, harsh way, then 2) ask ChatGPT to soften the tone so that it is achieves a specified tone. Then I take any edits that I agree with and can stand behind.
- Designing an infographic to help explain a concept after trying to explain it to a student a couple of ways already. My prompt to ChatGPT “You are a helpful tutor for [type of class]. A student is struggling with [concept] in your course. They learn best with visual examples. Create a 3-step remedial learning plan for this student. Include visual examples.” Then I edit the content with my own expertise.
- Updating statistics about prejudice attitudes on my slides. These stats aren’t the core content of my lecture, but I want the most up to date data. My prompt: “I’m looking for data from organization like Pew or Gallop that shows a graph or figure tracking attitudes about homosexuality over time. Graphic that answers the question: has homosexuality become more accepted by US society over time?”
- Other phrases I include in my prompts. “Do you have any questions for me to help you do this?” “What else do you need to know?” And if I run out of prompts, I just try a different GenAI (e.g., Claude, Microsoft Copilot).
A fellow GLCA professor, Professor Lew Ludwig has amassed several great resources for faculty getting started using AI. I particularly recommend scrolling to the section on “crafting prompts” and “AI and the environment“. Also includes information on what ChatGPT is.
Ideas to try right away
These ideas come from workshops hosted in 2025 by the Council of Independent Colleges.
Brainstorming assessments
- Think of a specific concept you teach. Use your AI tool and ask it to brainstorm 5 creative ways to assess student understanding of that concept. (E.g. “Brainstorm 5 creative assessments for a college-level history course on the topic of Bacon’s Rebellion”.
- At this point we want to be pretty vague. Don’t worry about feasibility yet, just collect initial ideas in a separate document.
- Try Ethan Mollick trick; the next instruction is just “Make it better.” Let AI output whatever it wants. Give this instruction 4 or 5 times.
- Look at the post “Make it Better” output and compare it to the output from the first prompt. What are the differences? Note what you notice and what you think happened.
Personalized Learning Pathway Prompt
- Here’s your challenge: A student is struggling with [core concept] in your course. They learn best with visual examples.
- Task: Write a prompt for an AI to create a 3-step remedial learning plan for this student.
- Next, ask it to create a short, personalized study guide for this student.
- Remember to include Persona (e.g., ‘You are a helpful tutor’), Context (the course, the student’s struggle), Task (create a study guide), and Format (e.g., ‘with 3 steps and visual examples’).
- You can also add to your prompt “what information do you need from me to do this task?”
Syllabus Reflection (for students to do)
- Credit: Susan Ray from Delaware County Community College
- Goal: you will use OpenAI as a vehicle to get to know our course, reflect on your strengths as a student, and consider what you might find interesting or challenging in ENGL-1XX. You’ll upload or paste our course syllabus into an AI tool, introduce yourself using a personalized prompt, and reflect on the AI’s response.
- Step 1: Prepare your personalized prompt. Start by filling in the prompt below, be sure to include:
- Your general age and intended major (for academic intersts)
- Your reading and writing background (this example comes from an English class)
- What you’ve liked or disliked about past English classes
- A few words about your learning style or future goals
- Example: “I’m a __-year old student thinking of majoring in ____. In the past I’ve found English classes to be ____. When it comes to writing, I tend to ___. I usually enjoy reading ___, but I find reading ___ more challenging. As a student and learner, my strengths include ___, but sometimes I struggle with ___. Based on the attached ENGL-1XX syllabus and what I’ve shared about myself, what parts of this course will I likely enjoy, what might be difficult, and how might it help me grow as a thinker and writer? How might this class help me in my career in ____?”
Syllabus Policy Co-Creator & Assignment Check
- Goal: Use AI to collaboratively draft a nuanced AI policy for your syllabus, helping you to articulate your pedagogical goals and expectations for students. Also, guidelines for individual assignments based on this policy.
- 1. Think about a specific course you teach and your general stance on student AI use (e.g., prohibited, allowed for brainstorming, encouraged with citation).
- 2. In an AI tool, use the prompt: “Act as an experienced university instructor who is designing a student-centered syllabus. My course is [Your Course Title], and the main goal is for students to learn [Key Skill or Concept]. I want to create a policy on the use of generative AI. Draft a concise, one-paragraph policy that is [Permissive/Restrictive/Balanced] and clearly explains the reasoning behind the rule.”
- 3. Review the output. Is the tone right? Is the reasoning clear?
- 4. Refine it with a follow-up prompt: “That’s a good start. Now, add a bulleted list of 3 specific examples of what constitutes ‘acceptable use’ and ‘unacceptable use’ according to this policy.”
- 5. Ask for a final touch: “Rephrase the entire policy to have a more encouraging and less punitive tone, framing it as a guide to academic integrity rather than a list of restrictions.”
- 6. Then, based on the policy you produce, ask the AI the following: “Based on this policy, review the assignments and assessments in the course. Use a Red-Yellow-Green system that could help students understand when it is acceptable and not acceptable to use AI in this course. Explain your reasoning to me about why you chose the color designation and explain what that designation means for the specific assignment to the student.”
Assignment design in age of AI
On AI-proof or AI-Aware assignments, this discussion “The Homework Machine” from the TeachLab podcast covers redesigning an assignment with AI in mind.
The Un-cheatable Assignment: This article explores designing assignments that are “un-cheatable” in the age of AI by requiring higher-order thinking skills like creativity and critical analysis. The author argues that instead of trying to prevent AI use, educators should create tasks that leverage AI as a learning tool. Ultimately, the essay concludes that this shift fundamentally changes the role of the educator from a content provider to that of a coach, mentor, and guide who helps students navigate complex problems with new tools.
See also the ideas in ‘Designing “AI-resistant” assignments‘ above.
AI-for-Equity ideas
Some faculty think about incorporating GenAI as a way to foster equity in the classroom. You might use it as an equalizer or support tool for students with diverse needs, especially those who may struggle with expressing themselves through writing (e.g., students with learning disabilities, English language learners). Other articles discuss using AI to build confidence with learning, not as a substitution for learning (e.g., AI as an optional tool for scaffolding, like brainstorming, outlining, or peer-review of a draft).
| Idea | What it does | How you might try it |
|---|---|---|
| 1. Offer AI-based assistive tools for students reading/writing or other challenges | AI-powered text-to-speech, speech-to-text, spell checkers, writing support can reduce barriers for students with learning differences or when English is not a first language | On a syllabus or Moodle site, suggest optional tools (e.g., text-to-speech in internet browsers or through K’s Level-Up-Learning) for students who prefer or need them; mention that using such tools is allowed and supported. |
| 2. Design assignments with an AI-option | By giving students the option to use AI (or not), you support diverse learner needs and preferences (i.e., Universal Design for Learning) and avoid penalizing those without access or those who dislike AI | Structure assignments as “AI-friendly but optional”; e.g., allow for brainstorming, outlining, reviewing rough drafts. You can give students specific prompts to use with GenAI to make sure they’re within the bounds you set. |
| 3. Explicitly talk about AI, bias, and digital equity; integrate AI literacy into your course | Because AI systems can reproduce biases or inequities, teaching about AI’s limits, potential biases, and ethical use helps students (and faculty) become critical & aware users. | Set aside time for a discussion or activity early in class about AI, equity, accessibility, bias, and digital divides. You may be surprised by what many students already know and think about. You can find many resources to help you guide your conversation. I like this one on cake. |
Neurodiversity
The GLCA also recently hosted a virtual workshop (recorded) called How Neurodiverse Students Can Use AI to Enhance Their Learning with Dr. Todd Zakrajsek.
AI-Creative Resources
Ideas to try
These ideas come from workshops hosted in 2025-26 by the Council of Independent Colleges (CIC).
50-Day AI Plan for Teaching
- This 50-day plan is designed to help you build fluency with AI through short, daily exercises of 15-30 minutes each.
- You can use it by either following the specific tasks provided or by adapting the core idea to something you are already working on for that day.
- The plan is broken down into phases that build your skills over time. As you move through the subsequent phases, the exercises will become more complex, helping you to leverage AI for more advanced pedagogical strategies and streamline your workflow.
- The ultimate goal is that after 50 days, you will intuitively see new possibilities for how AI can support you in your teaching.
Build a customized bot: Hyper-personalized learning
- CIC hosted a workshop series on Hyper-Personalized Learning (February 5) that explored how faculty can use customized chatbots to support teaching and learning—without replacing the human work at the center of education.
- Core framing of the workshop
- GenAI is positioned as an accompanying guide, not a substitute for instructors or peer interaction
- The goal is to augment the teaching and mentoring faculty already do, providing students with more 1-on-1 guidance personalized to their work
- Emphasis is on learning and thinking, not just producing answers
- Overview
- The series centered on how AI can tailor learning experiences to individual needs through adaptive content, flexible pacing, and real-time formative feedback, while maintaining a strong emphasis on inclusivity and accessibility. Speakers demonstrated practical use cases, such as AI tutors embedded in LMS platforms that provide precise guidance grounded in course resources, and 24/7 feedback assistants that help students iterate before submission—expanding timely support without replacing instructors’ expertise.
- Example highlighted in the session
- Create a customized chatbot that runs a case study
- Engage students in a back-and-forth learning dialogue
- Prompt students to explain reasoning, consider evidence, and revise thinking
- Instructors can review chatbot transcripts; assess student learning processes, not just final answers
- Want to try it yourself?
- If you’re ready to experiment can start with BoodleBox’s guide to building customized chatbots.
- Here is CIC’s guide to building effective bots
The AI-Powered Project Designer
- Goal: Brainstorm a creative and engaging student project where the use of generative AI is not only allowed but a required and integral part of the process.
- 1. Choose a major unit or module from one of your course.
- 2. Prompt the AI: “Act as an innovative curriculum designer for higher education. I want to create a final project for my [Your Course Topic] module. The project must require students to use a generative AI tool as a key part of their process. Brainstorm three distinct project ideas that go beyond simply writing an essay.” (BONUS: Use an assignment you already have and ask the AI to transform it by making a facet of the assignment use AI by necessity.)
- 3. Review the ideas. Are they creative? Are they aligned with your learning goals?
- 4. Select the most promising idea and ask the AI to flesh it out: “The [Chosen Project Idea] is the most interesting. Please create a brief project description for the syllabus. Include: 1) a short overview, 2) a list of three learning objectives, and 3) a sentence explaining how the use of AI will be assessed.”
Leveraging AI Tools for your Research
- Recently GLCA hosted an AI workshop by Lew Ludwig on using AI to power your own research agenda. From the talk: “Think of AI as a drafting partner/creative assistant — it can help organize, clarify, and even push your thinking forward. But you’re still the lead author. The voice, the judgment, the responsibility? That’s yours.”
- The workshop was recorded, a copy of the handout is included, and you can email Brittany Liu for a copy of the slides.
The “Pain Point” Blueprint – getting started with your own ‘inquiry guide’
- Goal: Design your specific “Inquiry Guide.”
- Step–by-Step:
- Identify: Where do students crash? (e.g., “Thesis Statements”).
- Personify: Who helps them? (e.g., “The Skeptic”).
- Constrain: Block the easy path. (e.g., “Never write the thesis.”).
- Ask for Help: Ask the AI: “I want to build a bot that helps students with X but never
gives the answer. What System Prompt should I use?” - Blueprint: Write down your trio: Pain Point + Persona + Constraint.
Exploring pedagogical uses of AI chatbots
- Exploring the pedagogical uses of AI chatbots – This guide from Stanford’s Teaching Commons explores the various pedagogical uses of AI chatbots to enhance teaching and learning.
- It outlines seven key approaches for integrating AI, including using chatbots as a mentor, tutor, coach, teammate, student, simulator, or tool to support a wide range of educational activities.
- To get started, the page offers practical advice on crafting effective prompts and provides hands-on exercises to help you practice and discover the capabilities of these AI tools.